How S&P Global Hit 15-Minute RTO on Capital IQ Using Amazon FSx and NetApp ONTAP
Cloud Platform EngineeringFinancial data platforms can’t afford the recovery time most enterprise disaster recovery architectures assume. When S&P Global Market Intelligence needed to protect Capital IQ from a regional AWS failure, they didn’t just replicate their data—they pre-provisioned their recovery environment so failover becomes a traffic redirect rather than a rebuild.
What Happened
S&P Global deployed a two-tier disaster recovery architecture on AWS using Amazon FSx for NetApp ONTAP, SnapMirror replication, and a Windows Server Failover Cluster spanning US-East-1 and US-West-2. SnapMirror runs on 15-minute schedules with actual RPO dropping to a few minutes during lower-activity periods. The key structural innovation is FlexClone pre-provisioning: a read-only instance is cloned daily during idle periods, eliminating the infrastructure build step from the disaster recovery sequence.
During a regional failure, the recovery sequence becomes: redirect traffic to the DR region, where the cloned read-only environment is already running. FlexClone creation itself completes in under two minutes. The result is sub-15-minute RTO for read-only access, with full read-write recovery following through the WSFC layer spanning both regions.
Why It Matters
Most enterprise DR architectures are built to recover from failure. This one is built to have already recovered before failure occurs. The distinction matters for financial data platforms with regulatory and contractual uptime obligations. Traditional recovery—standing up infrastructure during the incident window—adds 30–60 minutes to any realistic RTO even with well-rehearsed runbooks. FlexClone pre-provisioning shifts that build time to idle periods and converts the disaster scenario from a construction problem to a switching problem.
The SnapMirror-to-FlexClone pattern is available to any organization running Amazon FSx for NetApp ONTAP. S&P Global’s architecture serves as a validated enterprise reference for the capability published directly on the AWS Architecture Blog. The four-node WSFC spanning two regions provides the network layer for seamless failover without application-level reconfiguration.
What This Means for Platform Teams
Platform teams with critical data workloads should audit their current DR RTO against what FSx for NetApp ONTAP makes possible. The gap between a 15-minute pre-provisioned failover and a 60-minute traditional rebuild is not a marginal optimization—it’s the difference between a service disruption and a service outage from a customer and regulatory perspective. Teams running AWS-based high-availability workloads should evaluate whether pre-provisioned cloning fits their RPO/RTO targets before the next DR exercise surfaces the rebuild-time gap.
The So What
S&P Global’s case establishes that sub-15-minute RTO is achievable at enterprise scale for regulated financial data using existing AWS-native capabilities. The SnapMirror-to-FlexClone design pattern is the mechanism worth internalizing: pre-provision the recovery environment during idle periods so failover is a switch, not a construction project. Any platform team operating under SLA pressure should benchmark their current DR architecture against this reference before the next incident window tests it.
Source: AWS Architecture Blog
Content created with AI assistance and reviewed for accuracy.
Join the conversation
Stack Insiders is our free community for readers who want to go deeper — share resources, ask questions, and connect with others across every vertical we cover.
Join Stack Insiders →